Skip to content

tfidf runtime enhancement changes #1571

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
miguelgfierro merged 3 commits into from
Dec 3, 2021
Merged

tfidf runtime enhancement changes #1571

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
810032e
tfidf runtime enhancement changes
AdityaSoni19031997 Nov 30, 2021
fd94c95
making the snip flake8 compatible.
AdityaSoni19031997 Nov 30, 2021
82010a5
removing variable_name prefixed with an underscore
AdityaSoni19031997 Dec 2, 2021
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
33 changes: 17 additions & 16 deletions recommenders/models/tfidf/tfidf_utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -68,10 +68,12 @@ def __clean_text(self, text, for_BERT=False, verbose=False):
clean = clean.replace("Â\xa0", "") # non-breaking space

# Remove all punctuation and special characters
clean = re.sub("([^\s\w]|_)+", "", clean) # noqa W695 invalid escape sequence '\s'
clean = re.sub(
r"([^\s\w]|_)+", "", clean
) # noqa W695 invalid escape sequence '\s'

# If you want to keep some punctuation, see below commented out example
# clean = re.sub('([^\s\w\-\_\(\)]|_)+','', clean)
# clean = re.sub(r'([^\s\w\-\_\(\)]|_)+','', clean)

# Skip further processing if the text will be used in BERT tokenization
if for_BERT is False:
Expand Down Expand Up @@ -239,13 +241,17 @@ def __create_full_recommendation_dictionary(self, df_clean):
# Similarity measure
cosine_sim = linear_kernel(self.tfidf_matrix, self.tfidf_matrix)

# sorted_idx has the indices that would sort the array.
sorted_idx = np.argsort(cosine_sim, axis=1)

data = list(df_clean[self.id_col].values)
len_df_clean = len(df_clean)

results = {}
for idx, row in df_clean.iterrows():
similar_indices = cosine_sim[idx].argsort()[: -(len(df_clean) + 1) : -1]
similar_items = [
(cosine_sim[idx][i], df_clean[self.id_col][i]) for i in similar_indices
]
results[row[self.id_col]] = similar_items[1:]
for idx, row in zip(range(0, len_df_clean), data):
similar_indices = sorted_idx[idx][: -(len_df_clean + 1) : -1]
similar_items = [(cosine_sim[idx][i], data[i]) for i in similar_indices]
results[row] = similar_items[1:]

# Save to class
self.recommendations = results
Expand All @@ -264,17 +270,12 @@ def __organize_results_as_tabular(self, df_clean, k):
rec_item_id = list()

# For each item
for idx in range(0, len(self.recommendations)):
for _item_id in self.recommendations:
# Information about the item we are basing recommendations off of
rec_based_on = list(self.recommendations.keys())[idx]
tmp_item_id = str(
df_clean.loc[df_clean[self.id_col] == rec_based_on][self.id_col].values[
0
]
)
rec_based_on = tmp_item_id = _item_id

# Get all scores and IDs for items recommended for this current item
rec_array = self.recommendations[rec_based_on]
rec_array = self.recommendations.get(rec_based_on)
tmp_rec_score = list(map(lambda x: x[0], rec_array))
tmp_rec_id = list(map(lambda x: x[1], rec_array))

Expand Down